Cognitive Linguistics 16.02.2004 PT Part Three LEAVE NO VERB BEHIND (Psycholinguistic) research on categorization by prototype originated from work on.

Cognitive Linguistics 16.02.2004Cognitive Linguistics 16.02.2004

PT Part ThreeLEAVE NO VERB BEHIND

(Psycholinguistic) research on categorization (Psycholinguistic) research on categorization by prototype originated from work on colour by prototype originated from work on colour extends to less perceptual related extends to less perceptual related categories, concrete natural or nominal categories, concrete natural or nominal concepts (e.g. birds, furniture)concepts (e.g. birds, furniture)

? What about verbs? ? What about verbs? What kind of concepts are expressed by verbs?What kind of concepts are expressed by verbs?Are verbal concepts organized the way nominal Are verbal concepts organized the way nominal

concepts are?concepts are? Illustration: a corpus-linguistic approach to Illustration: a corpus-linguistic approach to

nine near synonyms that mean nine near synonyms that mean try try in in RussianRussian

We will go overWe will go over

Taylor (1995) ch. 3 (& 4): Taylor (1995) ch. 3 (& 4): Classical approach vs prototype approach Classical approach vs prototype approach

to categorizationto categorization Necessary & sufficient conditions Necessary & sufficient conditions

prototypesprototypes

– Zoom in: Horizontal category structure: Zoom in: Horizontal category structure: the protypethe protype

– Zoom out: Vertical category structure: Zoom out: Vertical category structure: the basic levelthe basic level

Aristoteles vs WittgensteinAristoteles vs Wittgenstein

– Aristotelian view on categories ≠ Aristotelian view on categories ≠ psychologically real psychologically real inadequacy of inadequacy of necessary and sufficient attributesnecessary and sufficient attributes

– Binary features Binary features family resemblance, family resemblance, overlapping clusters overlapping clusters

– Category boundaries: strict Category boundaries: strict can be can be fuzzyfuzzy

– Members: equal Members: equal degrees of degrees of representativenessrepresentativeness

! ! These four properties need not cooccur (Lakoff These four properties need not cooccur (Lakoff 1987: 56, Geeraerts 1989: 225)1987: 56, Geeraerts 1989: 225)

Horizontally: the Prototype Horizontally: the Prototype (1975)(1975)

Rosch&Mervis 1975: 515 (Taylor 1995: 59)Rosch&Mervis 1975: 515 (Taylor 1995: 59)

Prototype = the abstract cluster possessing the Prototype = the abstract cluster possessing the characteristics with the highest cue validity for characteristics with the highest cue validity for the category. This abstract cluster can manifest the category. This abstract cluster can manifest itself in a particular entity that exhibits these itself in a particular entity that exhibits these highly typical and important attributeshighly typical and important attributes

Cue validity is defined in terms of its total frequency Cue validity is defined in terms of its total frequency within a category and its proportional frequency within a category and its proportional frequency in that category relative to contrasting categories in that category relative to contrasting categories = t= the prototype is very similar to other members he prototype is very similar to other members of the same category and very dissimilar to of the same category and very dissimilar to members of other categoriesmembers of other categories

Prototype = (1995)Prototype = (1995)

Hampton (1995: 104-105): a prototype concept has Hampton (1995: 104-105): a prototype concept has three propertiesthree properties

1.1. the prototype representation itself: a the prototype representation itself: a generalisation or abstraction of some central generalisation or abstraction of some central tendency, average or typical value of a class of tendency, average or typical value of a class of instances falling in the same category. instances falling in the same category.

2.2. the prototype concept specifies a way of defining the prototype concept specifies a way of defining similarity to this prototype. similarity to this prototype.

3.3. there is a criterion or cut-off level of similarity for there is a criterion or cut-off level of similarity for category membership. Passing the similarity category membership. Passing the similarity criterion is both necessary and sufficient for criterion is both necessary and sufficient for category membership. ! lack of necessary and category membership. ! lack of necessary and sufficient conditions ≠ there are no membership sufficient conditions ≠ there are no membership criteria at all. criteria at all.

An efficient way of An efficient way of structuringstructuring

(Geeraerts 1985: 141)(Geeraerts 1985: 141) requirments posed on cognitionrequirments posed on cognition

– Structural stability: deal with atypical Structural stability: deal with atypical and damaged examples of a categoryand damaged examples of a category

– Flexible adaptability: changing Flexible adaptability: changing conditions ~ changing expressive needsconditions ~ changing expressive needs

Vertically: the basic levelVertically: the basic level

2 axes of categorization: an entity 2 axes of categorization: an entity may be categorized in different ways, may be categorized in different ways, on different levelson different levels

Hierarchical organization: economicHierarchical organization: economicThree levelsThree levels

SuperordinateSuperordinateBasic Basic SubordinateSubordinate

Category type Gestalt Attributes Structure Function Linguistic form

Sub-ordinate Almost identical Gestalt

Large number category-wide attributes; salient specific attribute(s)

High degree of homogeneity among category members

Specify-ing function

Often morphologi-cally complex words

Basic level Common Gestalt

Large number of category-wide attributes

Prototype structure

“natural” access to the world

Short, one- morphemic words

Super-ordinate

No common Gestalt

One or very few category-wide attributes; salient general attribute(s)

Family resemblance structure

Highligh-ting and collecting function

Often longer, morphologi-cally complex words

What about non-nouns?What about non-nouns? Verbs: Verbs:

– Coleman and Kay (1981) on Coleman and Kay (1981) on lielie– Fillmore (1982) on Fillmore (1982) on climbclimb– Pulman (1983): verbs of killingPulman (1983): verbs of killing– Geeraerts (1985/1988) on Geeraerts (1985/1988) on vernietigen/vernielenvernietigen/vernielen and and

verduwen/verdouwenverduwen/verdouwen– Schmid 1993 on Schmid 1993 on begin/startbegin/start

AdjectivesAdjectives– Taylor (2003) on Taylor (2003) on high/tallhigh/tall

PrepositionsPrepositions– Brugman (1981) on Brugman (1981) on over over

= polysemous and synonymous items= polysemous and synonymous items

What are near-synonymsWhat are near-synonyms

Cruse 1986 & 2000Cruse 1986 & 2000

Absolute, perfect or full synonyms “Absolute, perfect or full synonyms “if and only if if and only if all contextual relations are identical”all contextual relations are identical” non- non-existentexistent

Cognitive synonyms “must be identical in respect Cognitive synonyms “must be identical in respect of propositional traits, but may differ in respect of of propositional traits, but may differ in respect of expressive traitsexpressive traits eg eg vviolin/fiddleiolin/fiddle

Plesionyms yield sentences with different truth-Plesionyms yield sentences with different truth-conditions: weak contrast, permissible differences conditions: weak contrast, permissible differences must be either minor or backgroundedmust be either minor or backgrounded

Foregrounded minor distinctionsForegrounded minor distinctions– scales of degree: scales of degree: fogfog and and mistmist– adverbial specializations of verbs: adverbial specializations of verbs: ambleamble (or (or

walking ???) vs walking ???) vs strollstroll (or walking ???) (or walking ???)– aspectual distinctions: the state of being aspectual distinctions: the state of being calmcalm

and the predisposition of being and the predisposition of being placidplacid – differences of prototype center: focus on differences of prototype center: focus on

physical factors for physical factors for bravebrave vs focus on vs focus on intellectual and moral factors in intellectual and moral factors in courageouscourageous. .

Backgrounded major distinctions like Backgrounded major distinctions like gender can be found in the pair gender can be found in the pair prettypretty (for (for women) and women) and handsomehandsome (for men) (for men)

Some pairs can be “more synonymous” Some pairs can be “more synonymous” than other pairs than other pairs scale of synonymy scale of synonymy framework that can deal with thisframework that can deal with this

PT and near-synonymsPT and near-synonyms

PT seems ideally equipped to deal with PT seems ideally equipped to deal with near-synonymsnear-synonyms– Family resemblance structure of clustering and Family resemblance structure of clustering and

overlapping sensesoverlapping senses Share greater part of their meaningShare greater part of their meaning Resemble each other in different respectsResemble each other in different respects

– Gradedness & requirment of sufficient similarityGradedness & requirment of sufficient similarity Belong to a category only to a certain degree, be a Belong to a category only to a certain degree, be a

better or worse example of a conceptbetter or worse example of a concept

– Fuzzy boundariesFuzzy boundaries Difficult to draw the line: compare entries in Difficult to draw the line: compare entries in

dictionaries dictionaries

PT and near synonymous PT and near synonymous verbsverbs

Challenges both on the level of vertical (1/2) Challenges both on the level of vertical (1/2) and horizontal (3) structuring (Pulman 1983)and horizontal (3) structuring (Pulman 1983)– Taxonomies (‘is a’) for verbs are artificial constructs Taxonomies (‘is a’) for verbs are artificial constructs

partonomies (‘is a part of’), scripts, frames … partonomies (‘is a part of’), scripts, frames …– Near synonyms are situated on the subordinate Near synonyms are situated on the subordinate

level where prototypes are less well-developedlevel where prototypes are less well-developed– Prototypes for verbs necessarily refer to more Prototypes for verbs necessarily refer to more

complex properties than perceptual or functional complex properties than perceptual or functional onesones

? Rather similarity to category name than to each other? Rather similarity to category name than to each other– Kill: murder, strangle, execute, assassinateKill: murder, strangle, execute, assassinate– Walk: wander, march, stride, strollWalk: wander, march, stride, stroll– ???: try, attempt, endeavour???: try, attempt, endeavour

basic experiences instead of basic levels (Ungerer basic experiences instead of basic levels (Ungerer & & Schmid 1996) ~ ICMs (Lakoff 1987)Schmid 1996) ~ ICMs (Lakoff 1987)

PT and near synonymous PT and near synonymous verbsverbs

Challenges related to type of concept expressed Challenges related to type of concept expressed – Verbs express relatively abstract conceptsVerbs express relatively abstract concepts

no intuitively obvious hierarchy of sets into which no intuitively obvious hierarchy of sets into which such concepts can be placedsuch concepts can be placed (Hampton 1981)(Hampton 1981)

membership of abstract concepts is almost membership of abstract concepts is almost unlimited in the range of possible new instances unlimited in the range of possible new instances that could be discovered or inventedthat could be discovered or invented

- Prototype theory only applies indirectly to - Prototype theory only applies indirectly to lexical lexical categories, through the role it plays categories, through the role it plays in in understanding the structure of their understanding the structure of their signifieds signifieds (Vandeloise 1990: 435)(Vandeloise 1990: 435)

verbs like verbs like try try express relational concepts, they express relational concepts, they do not stand do not stand alone and cannot be handled by looking at alone and cannot be handled by looking at the verb in isolationthe verb in isolation

the prototype for abstract concepts is reduced the prototype for abstract concepts is reduced to the way to the way the uses of the concepts are organized into the uses of the concepts are organized into “prototypical “prototypical structures” (Geeraerts 1985) or “central structures” (Geeraerts 1985) or “central use” (Rachilina 2000)use” (Rachilina 2000)

Organizing better and worse Organizing better and worse exemplarsexemplars

internal strucure of a category that is internal strucure of a category that is organized by a prototype = ?organized by a prototype = ?

Complex categories ~ networks of Complex categories ~ networks of interrelated entities/senses/…interrelated entities/senses/…– Schematic (Langacker 1987): full Schematic (Langacker 1987): full

sanctionsanction– Radial (Lakoff 1987): partial sanctionRadial (Lakoff 1987): partial sanction

They are compatibleThey are compatible

Schematic vs Radial Schematic vs Radial NetworksNetworks

TREE

Oak Elm Maple

Birth mother

Adoptivemother

Fostermother

Mother

Three problems of Three problems of synonymysynonymy

? ? Delineation: which verbs should be Delineation: which verbs should be mentioned together and which ones mentioned together and which ones should be left outshould be left out

? Structuring: do synonym-clusters ? Structuring: do synonym-clusters have an internal structure and what have an internal structure and what could/does it look likecould/does it look like

? Description: is there a verifiable way ? Description: is there a verifiable way to describe the scales of variation to describe the scales of variation along which the synonyms differalong which the synonyms differ

CL “difference in form entails CL “difference in form entails difference in meaning”difference in meaning”

Formal differences can be used to Formal differences can be used to delineate and structure clusters of delineate and structure clusters of near synonyms near synonyms – Are all verbs created equal? Are all verbs created equal? – How ‘near’ is ‘near’? How ‘near’ is ‘near’?

Formal differences provide Formal differences provide interesting data for the scales of interesting data for the scales of variation within clustersvariation within clusters

Case study on near synonyms Case study on near synonyms expressing expressing trytry in Russian in Russian

Are all verbs created equal? Are all verbs created equal? A constructional view.A constructional view.

CL: “humans have the capacity to CL: “humans have the capacity to construe conceptual content in construe conceptual content in alternate ways”alternate ways” (Langacker 1990: 5) (Langacker 1990: 5)

Constructions convey meaning at the Constructions convey meaning at the coarse grained levelcoarse grained levelVariation along 3 constructional Variation along 3 constructional

parametersparameters alternations specify the degree of alternations specify the degree of

integration between the finite verb integration between the finite verb and the infinitiveand the infinitive

alternations encode different levels of alternations encode different levels of conceptual (inter)dependence conceptual (inter)dependence

Combination of constructional possibilities Combination of constructional possibilities continuum of decreasing conceptual continuum of decreasing conceptual in(ter)dependence for the finite verbin(ter)dependence for the finite verb

II II II III III IV IV V V VI VI VII VII VIIIVIII

+prop+prop +prop+prop +prop+prop +prop+prop -prop-prop -prop-prop -prop-prop -prop-prop

+comp+comp + comp+ comp - comp- comp - comp- comp + comp+ comp + comp+ comp - comp- comp - comp- comp

+temp+temp

/loc/loc-temp -temp /loc/loc

+temp+temp

/loc/loc-temp -temp /loc/loc

+temp+temp

/loc/loc-temp -temp /loc/loc

+temp+temp

/loc/loc-temp -temp /loc/loc

P-verbP-verb Semi Semi P-verbP-verb SemiSemi M-verbM-verb

M-verbM-verb

M-VERBS V IV III II I

Modal Intentional Tentative Resultative Phasal МочьIP (can)

ДуматьIP (think of, plan to) НамереватьсяIP (intend, mean to) ПредполагатьIP (intend to, contemplate) СобиратьсяIP (intend, be about) ХотетьIP (want, wish)

(ПоP)пробоватьIP (try) (ПоP)пытатьсяIP (try, attempt, endeavor) (ПоP)старатьсяIP (try, endeavour, seek) НоровитьIP (strive to, aim at) ПорыватьсяIP (try, endeavor) ПыжитьсяIP (go all out) СилитьсяIP (try, make efforts) ТужитьсяIP (make an effort) ТщитьсяIP (try, endeavour)

Не задуматьсяP (not to hesitate) (не) ПоспетьP (have time) (не) УдосуживатьсяIP/удосужитьсяP (find time to, manage) (не) УспеватьIP/успетьP (have time, manage) ИзловчитьсяP (contrive, manage) ИсхитритьсяP (manage, contrive) УмудрятьсяIP/умудритьсяP (contrive, manage) УхитрятьсяIP/ухитритьсяP (manage, contrive) СмочьP (be able to) СуметьP (be able to)

НачинатьIP/начатьP (begin, start, commence) СтатьP (begin, start) ПродолжатьIP (continue, go on) ПереставатьIP/перестатьP (stop, cease) ПрекращатьIP/прекратитьP (stop, cease, discontinue) БросатьIP/броситьP (give up) КончатьIP/кончитьP (finish, end) ЗаканчиватьIP/закончитьP (end, finish)

Evidence Evidence for the coherence of the categories for the coherence of the categories the composition of the verbal inflectional paradigm the composition of the verbal inflectional paradigm

(aspect, mode, tense, cf. Haiman 1985)(aspect, mode, tense, cf. Haiman 1985) the observation that each category of verbs the observation that each category of verbs

– prefers a specific aspect prefers a specific aspect ~ Vendlerian (1967) verbal ~ Vendlerian (1967) verbal ‘situation types’ & exercises a specific influence on ‘situation types’ & exercises a specific influence on the aspect of thethe aspect of the infinitiveinfinitive

– reacts homogeneously in implicational tests reacts homogeneously in implicational tests (Karttunen 1971)(Karttunen 1971)

Cross-linguistically valid notion M-verbsCross-linguistically valid notion M-verbs– Givón (1990: 517, 533): “modality verbs” Givón (1990: 517, 533): “modality verbs” – Dixon (1996): “secondary verbs”: not necessarily Dixon (1996): “secondary verbs”: not necessarily

verbal lexemes, semantic modification to core verbal lexemes, semantic modification to core concept (= infinitive)concept (= infinitive)

Delineation (+/-) solved Delineation (+/-) solved

= = 9 verbs: 9 verbs: пробовать, пытаться, стараться, пробовать, пытаться, стараться, силиться, тщитьсясилиться, тщиться, , порываться, норовить, порываться, норовить, пыжиться, тужитьсяпыжиться, тужиться

== convey similar meanings (‘try, attempt, convey similar meanings (‘try, attempt, endeavor’) in a similar way (exclusively endeavor’) in a similar way (exclusively +inf, specific linear distribution)+inf, specific linear distribution)

constructional & linear distributional data constructional & linear distributional data provide a way to determine whether two provide a way to determine whether two or more verbs are sufficiently similar to be or more verbs are sufficiently similar to be considered near-synonymsconsidered near-synonyms

+ verifiable & repeatable + verifiable & repeatable previous previous analysesanalyses

– Evgen’evaEvgen’eva (2001²): (2001²): пытаться, пробоватьпытаться, пробовать; ; стараться, стремиться, пытаться, норовить, стараться, стремиться, пытаться, норовить, силиться, тщиться, пыжиться силиться, тщиться, пыжиться [lacks порываться[lacks порываться and and тужиться]тужиться]

– ČernovaČernova (1996: 87): по/пытаться, тщиться, (1996: 87): по/пытаться, тщиться, домогаться, искать, не почесаться, домогаться, искать, не почесаться, по/стараться, силиться; по/стараться, силиться; попо//пробоватьпробовать; ; норовить норовить [lacks порываться, пыжиться[lacks порываться, пыжиться and and тужиться]тужиться]

– Apresjan Apresjan (1999²: 303-308): (1999²: 303-308): пытаться, пытаться, пробовать, стараться, силитьсяпробовать, стараться, силиться; [related but not ; [related but not near-synonymous]: near-synonymous]: добиваться; домогаться; добиваться; домогаться; хотеть, намереватьсяхотеть, намереваться; ; стремиться, рваться, стремиться, рваться, порыватьсяпорываться;; биться, осилить биться, осилить;; тщиться, тщиться, пыжитьсяпыжиться;; норовить норовить;; испытывать, пробовать2 испытывать, пробовать2 [lacks тужиться][lacks тужиться]

How ‘Near’ Is ‘Near’?How ‘Near’ Is ‘Near’?A corpus-based view on A corpus-based view on

internal structureinternal structure

Do synonym-clusters have an internal structure and Do synonym-clusters have an internal structure and what does it look like? ~ intuitions of what does it look like? ~ intuitions of lexicographerslexicographers meaning // conceptmeaning // concept unifying conceptual analysis // non-linguistic mental unifying conceptual analysis // non-linguistic mental

capacitiescapacities Prototype structure of conceptsPrototype structure of concepts Radial networkRadial network

Internal structure: ≠ clusters ≠ central conceptsInternal structure: ≠ clusters ≠ central concepts ‘‘Elaborative distance’ (Langacker 1987) ~ scale of Elaborative distance’ (Langacker 1987) ~ scale of

synonymy synonymy

How can we extract the information needed without How can we extract the information needed without carrying our psycholinguistic (elicitation) carrying our psycholinguistic (elicitation) experiments? experiments? textcorpus: non-elicited datatextcorpus: non-elicited data

Native speakers go through an unconscious Native speakers go through an unconscious decision process decision process choose one of the choose one of the possible verbs. possible verbs.

The factors that support the choice for a The factors that support the choice for a particular verb ~ the attributes of verbsparticular verb ~ the attributes of verbs

Frequency of mention in a corpus of texts Frequency of mention in a corpus of texts mirrors preference and degree of specificity: mirrors preference and degree of specificity: – variables that are mentioned more frequently variables that are mentioned more frequently

can be thought of as coming to mind first when can be thought of as coming to mind first when using a particular verb ~ typical attributesusing a particular verb ~ typical attributes

– Verbs that are mentioned more frequently and Verbs that are mentioned more frequently and with more variables fit more situations ~ with more variables fit more situations ~ prototypical usageprototypical usage

SubjectSubject: case (nom, acc, dat) + type of subject : case (nom, acc, dat) + type of subject (human, animal, insect vs. concrete (artificial, (human, animal, insect vs. concrete (artificial, natural, …) and abstract (concepts, organizations) natural, …) and abstract (concepts, organizations) objects …)objects …)

Finite verbFinite verb: aspect (impf/pf), mode (inf, ind, imp, : aspect (impf/pf), mode (inf, ind, imp, cond/subj, participle, gerund), tense (past, cond/subj, participle, gerund), tense (past, present, future)present, future)

InfinitiveInfinitive: aspect + type of action (physical (literal : aspect + type of action (physical (literal (5) vs figurative (3)), perceptual (2), (5) vs figurative (3)), perceptual (2), communicative, mental, emotional ~ CL communicative, mental, emotional ~ CL ‘‘embodiment principle’, ‘basic domains’) + degree embodiment principle’, ‘basic domains’) + degree of controlof control

Collocates within the same clauseCollocates within the same clause: esp. : esp. adverbs (duration, intensity, repetition, vainness), adverbs (duration, intensity, repetition, vainness), particles (e.g. negation) and connectorsparticles (e.g. negation) and connectors

Clause typeClause type: main vs subclause: main vs subclause, declarative vs , declarative vs imperative vs interrogativeimperative vs interrogative

Need some help Need some help processing?processing?

= 92 ID tags for 9 verbs in 1525 = 92 ID tags for 9 verbs in 1525 example sentences = 140 300 data example sentences = 140 300 data pointspoints

How to determine the precise degree of How to determine the precise degree of similarity between those verbs?similarity between those verbs?

How to determine where exactly to connect How to determine where exactly to connect a verb to another verb in the network?a verb to another verb in the network?

statistical techniquesstatistical techniquesCluster analysis: Cluster analysis: organize huge amounts of organize huge amounts of

observed data into meaningful structuresobserved data into meaningful structures (cf. (cf. Gries forthc.)Gries forthc.)

= = measure the similarity of the verbs in terms of measure the similarity of the verbs in terms of their ID tags / behavioral profiletheir ID tags / behavioral profile

Translate dendrogram into radial network Translate dendrogram into radial network // human categorization mechanisms// human categorization mechanismsinternal structure: 5 clusters of conceptually internal structure: 5 clusters of conceptually

distinct subcategories of near-synonyms distinct subcategories of near-synonyms

~ 5 Idealized Cognitive Models (Lakoff ~ 5 Idealized Cognitive Models (Lakoff 1987: 68, Hirst 1995) [+ implications]1987: 68, Hirst 1995) [+ implications]

Situation triggers affinities with one of Situation triggers affinities with one of the ICMs the ICMs that clusters is chosen that clusters is chosen

Internal structuring (+/-) Internal structuring (+/-) solvedsolved

Make explicit the “scales of variation of specific Make explicit the “scales of variation of specific properties along which neighboring items would be properties along which neighboring items would be plesionyms and distant ones non-synonyms, with plesionyms and distant ones non-synonyms, with possibly an indeterminate zone in between” (Cruse possibly an indeterminate zone in between” (Cruse 1986: 288): compare the verbs in one cluster and 1986: 288): compare the verbs in one cluster and perform clusteranalysis on a selection of tagsperform clusteranalysis on a selection of tags

type of attempt (aspect, mode, tense; clause type of attempt (aspect, mode, tense; clause type; adverbs): type; adverbs):

– repeated, repetitive-durative, durativerepeated, repetitive-durative, durative– intensiveintensive

type of infinitive action type of infinitive action (mode, control, negation, (mode, control, negation, particles): particles):

– control, reduced control, no control control, reduced control, no control chances chances of success (adverbs, connectors): low, medium, highof success (adverbs, connectors): low, medium, high

reasons for failure (connectors): external, reasons for failure (connectors): external, internalinternal

elaborative distance: ‘near’ can be far …elaborative distance: ‘near’ can be far …

Description (+/-) solvedDescription (+/-) solved

RecapitulationRecapitulation Difference in form entails difference in meaning Difference in form entails difference in meaning

– constructional differences can be linkedconstructional differences can be linked to conceptual differences: conceptually (inter)dependent to conceptual differences: conceptually (inter)dependent

verbsverbs

delineate the conceptual space of delineate the conceptual space of trytry

– constructional differences can be enriched withconstructional differences can be enriched with precise colligational and collocational information: ID tagsprecise colligational and collocational information: ID tags

– formal properties: 5 different formal properties: 5 different clusters or clusters or ICM’s for “to ICM’s for “to try” : show how the conceptual space is structured try” : show how the conceptual space is structured

compare verbs in one cluster to find the differences in compare verbs in one cluster to find the differences in prototype center [! Implications]prototype center [! Implications]

– (simplified) radial network representation ~ scale (simplified) radial network representation ~ scale ~ intuitions of native speakers-lexicographers~ intuitions of native speakers-lexicographers